We introduce grounded glossary generation, a structured task requiring models to recover semantically meaningful Sanskrit phrases and produce translation-grounded meanings from a sloka-translation pair, formalizing the traditional patha commentary practice as an evaluable NLP objective. We construct a benchmark of 31,316 sloka-translation-glossary triples from the Valmiki Ramayana and Srimad Bhagavatam, paired with two metrics: Jaccard for phrase recovery and Meaning Faithfulness for semantic consistency. Across zero-shot, few-shot, and instruction fine-tuned variants of Gemma-3n-E4B, Gemma-3-12B, Phi-4, and Qwen3.5-9B, instruction fine-tuning substantially outperforms prompting, while explicit segmentation yields gains. Error analysis identifies over-segmentation of sandhi and samasa compounds as the dominant failure mode, pointing to morphological modeling as the key bottleneck for faithful Sanskrit lexical decomposition.
We construct a corpus of 1,262 verse--commentary (urai) pairs from five Classical Tamil source sections, ranging from technical grammatical prose to modern paraphrase, and ask what information representation learning can recover. We train recurrent and Transformer encoders, a Siamese-style pair-matching network, an mBART-style encoder--decoder, and a decoder-only language model. Each analysis is interpreted against an appropriate control on the same data. TF-IDF provides a strong no-training lexical retrieval baseline, alongside representation analyses and generation controls for the learned models. A fixed string containing the 25 most frequent commentary words scores higher on generation overlap than the decoder-only model. Canonical correlation reaches 1.000 on Gaussian noise at these sample sizes, token-F1 spans only about 0.02--0.20 on this corpus, and the encoder--decoder continues to lower training loss for sixteen epochs after validation loss has begun to rise. One narrow result remains: the decoder-only model prefers authentic word order in 107 of 112 minimal-pair comparisons (95.5%), but does not reproduce held-out commentary content. We release the extraction and evaluation protocol; redistribution of the source commentaries remains subject to permission.
Tibetan-to-English machine translation (MT) models frequently falter under extreme domain data scarcity, often producing translations that violate the distinctive agglutinative rules of Tibetan and suffer from domain-specific stylistic mismatches. To overcome these limitations, we propose Semantic-Syntax Prealignment (SSPA), an innovative corpus generation framework. SSPA constructs high-quality pseudo-parallel pairs by explicitly minimizing the deviation between the syntactic-semantic profiles of generated samples and professional reference texts. Specifically, source-target structural representations are standardized through length-unified truncation and terminology normalization, followed by a dual-domain alignment process that maximizes syntactic cosine similarity under rigorous structural constraints. We further augment these aligned frames via a cross-length dynamic filling mechanism, which is integrated with an Expectation-over-Transformation (EOT)-based style regularization mechanism specifically adapted for stylistic perturbations, to simulate authentic linguistic variations. Extensive evaluations on our newly constructed Tibetan Medicine-Tibetan English (TM-TE) dataset demonstrate that SSPA significantly outperforms existing competitive baselines. Notably, SSPA achieves a BLEU-4 score of 36.2 and improves long-sentence BLEU-4 by 16.8 points, with a parser-verified grammatical compliance rate of 96.2%. The framework exhibits remarkable cross-domain adaptability and stylistic consistency, offering a robust, versatile solution for low-resource Tibetan professional domain MT.
Recent calls for harder machine translation benchmarks have not clarified what difficulty should mean. We argue that one meaningful and currently unmeasured axis is referential reach, the distance a segment must look back into its document to resolve the entities and pronouns it contains. We formalize this as discourse dependency (DDP), a metric-free, source-side measure computed from named entity re-mentions and pronominal coreference. Validated against gold coreference, DDP errs one-sidedly in 99.2% of segments, so a high-DDP segment is certified to require long-range context. Applying DDP to WMT24++ and WMT25 shows that both are heavily skewed toward low-DDP segments, which domain labels do not distinguish. Building on DDP, we compare five context injection strategies in an English-Korean post-editing setup, varying context size and selection. As DDP grows, no strategy keeps pace with human post-editing. On segments with DDP>= 15 raters prefer human translations, while automatic metrics register no difference. As frontier systems saturate aggregate scores, DDP shifts evaluation from how well models score to how far they can reach.
Ahrii Kim, Chanjun Park, Seong-heum Kim· 0 citations
Tamil spell and grammar correction is challenging because Tamil is an agglutinative low-resource language with rich verbal morphology, complex sandhi (phonetic transformation) rules at word boundaries, and a script of 247 distinct letters. Prior work targets word-level surface errors with rule-based methods, statistical n-gram models, Minimum Edit Distance, or hybrid pipelines with a transformer re-ranker; such methods cannot reliably handle contextual errors - subject-verb agreement, tense consistency, or cross-word sandhi - which require sentence-level understanding. We propose an end-to-end sequence-to-sequence formulation and fine-tune mT5-small and mBART-50 on a synthetic corpus of up to 657,720 noisy-clean Tamil sentence pairs spanning ten error categories. Both backbones follow the same four-stage progressive schedule, each stage targeting one weakness: surface noise (v2), contextual grammar (v3), single-site sandhi (v4), and multi-site cross-word sandhi (v5). On a 1,000-sentence balanced diagnostic set verified disjoint from all training data, our best model, mBART-50 v5, reaches 69.3% top-1 exact-match accuracy, with 87.5% on sandhi and 43.5% on subject-verb agreement. The schedule is what produces these gains: subject-verb accuracy rises from 1.0% to 52.5% once contextual pairs are introduced, and sandhi from 0% to 87.5% once multi-site sandhi pairs are. We additionally quantify a precision-recall trade-off this literature has not reported: sandhi recall is paid for monotonically in identity accuracy. Finally, Tamil-LLaMA-7B-Instruct reaches 19.0% zero-shot and 24.7% with three demonstrations against a 20.0% copy baseline, showing that a Tamil-adapted instruction model does not transfer to specialised sentence-level correction without task-specific supervision.
A. Karthikeyan, S. JayaNirmala, Sangeetha Sivanesan et al.· 0 citations
The exponential growth of scientific literature has intensified the demand for automated summarization systems capable of producing abstracts that are both linguistically fluent and factually reliable. Existing approaches face a fundamental trade-off: encoder-decoder models such as BART and T5 maintain strong factual grounding but produce rigid, extractive outputs, while decoder-only large language models (LLMs) such as Llama and Gemma generate highly fluent text yet remain susceptible to hallucination. This paper proposes a two-stage Synergistic Hybrid Ensemble framework designed to resolve this dichotomy. In Stage 1, a fine-tuned BART-Large model generates a factually grounded scaffold draft from a structured input representation comprising the document title, key sentences, method highlights, and results summary. In Stage 2, a QLoRA-adapted Llama-3.2-1B model performs coherent rewriting and stylistic polishing by conditioning on both the scaffold draft and the original source document. Experiments conducted on the arXiv Scientific Research Papers Dataset using BERTScore and entailment-based Factual Consistency metrics demonstrate that the proposed ensemble achieves a Factual Consistency metrics demonstrate that the proposed ensemble achieves a Factual Consistency score of 0.9140, substantially outperforming BART-Large (0.2890) and Llama-3.2-1B (0.6630) individually. Although the ensemble incurs a marginal reduction in BERTScore (0.8980) relative to Llama-3.2-1B (0.9555), this trade-off is justified given the critical importance of factual reliability in high-stakes scientific discourse. These findings confirm that anchoring the generative capacity of decoder-only LLMs to verified factual scaffolds effectively mitigates hallucination risk, offering a scalable and reproducible solution for high-fidelity scientific abstract generation.
Geoffrey Antonio Arifin, Andrew Widyanata, Henry Lucky et al.· International Conference on...· 0 citations
VIVID is presented, the first systematic benchmark for evaluating culturally grounded figurative language understanding in Vietnamese, and an evaluation framework combining generative and discriminative tasks, proposing an LLM-as-a-Judge approach with aspect-based prompting validated against human judgment.
Tu Tran Do, N. Nguyen, Tung Khanh Tran et al.· International Conference on...· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
